g2 6 Mr. Knox's experiments and observations 
The experiment was repeated, but without an attempt to 
catch the gases ; and the moment the smell came over the 
receiver was changed, by which means nearly the whole was 
obtained, independent of what it suffered from decomposition, 
and weighed about 2,5 grains per cent. 
Experiment 2. 
Finding that the iron decomposed the water, and thereby 
in part the bitumen also, I gave up that apparatus, and had a 
very strong glass retort, such as is used in the manufacture 
of common wine bottles, made of a similar shape, and coated, 
to which a receiver of the same form as those described was 
luted as before. 
I may mention here a very curious appearance of the mass 
which remained in the iron retort. It was of course of a 
cylindrical shape; its colour was pale ash-grey; it was po- 
rous, semivitrified, and slightly cohering ; resembling coarse 
pumice, and floated on water. But the circumstance to which 
I now allude is that it separated, when broken, into thin re- 
gular tables, precisely in the manner of madreporites, so that 
it might be considered as schistose pumice. 
In this experiment, I gave to 480 grains of the coarsely 
powdered leek-green variety, a dull red heat for half an hour, 
in order to expel the water, as I had found that, with that 
degree of heat, the bitumen was not removed from the stone. 
I then charged it into the glass retort, inserted in a crucible 
as before, and gave it a white heat. No gas was generated, 
nor expelled, and the liquor in the receiver was pure bitumen. 
It weighed 2,83 grains per cent, and was of a wine-yellow 
colour, and the smell the same as the last. The neck of the 
