304 PROCEEDINGS OF SOUIETIES. 
portions. The black globules are not all alike in composition ; 
some of these are silicates coloured by carbon, others are iron 
balls coated externally with a silicate. Many of these rusty cannon 
balls are probably ferrous oxide formed by the action of heat on 
the iron pyrites in the coal. There are also balls of black mag- 
netic oxide; the perforated shells are probably ferrous sulphides. 
The globular form of these bodies suggests that they have been 
thrown off in scintillations, such as are seen during the combus- 
tion of iron in oxygen gas, and whilst in a fluid state they assume 
a spheroidal form. They are carried by the draught into the flue, 
and being of greater specific gravity than the carbonaceous matter 
forming the smoke, they fall before the current of air has reached 
the chimney. Some of the dust has been a considerable time 
in the flue, exposed to the intensely heated circulating flame; 
the reducing action of this would probably convert some of the 
oxide into metalliciron. Many of these balls have the appearance 
of reduced oxides. The flue dust contains a larger amount of 
ferruginous matter than can be accounted for by the analysis of 
coal ash. I think the surplus may be regarded as representing 
the wear and tear of the iron work about the furnace, such as fire 
bars, boiler plates, &c. The brick work and cement about the 
boiler and flues may also supply some of the silica, alumina, 
and iron for these balls, numbers of which are merely thin 
shells. The movements of these objects, caused by the approach 
of a magnet under the stage of the microscope, are somewhat 
amusing, and it is at times startling to see the crystalline objects, 
both spherical and irregular, exhibit magnetic attraction: pro- 
bably they contain particles of iron imbedded in them; if they 
do not, may we not imagine that there is some magnetic com- 
pound in which the crystalline matter predominates? When 
we consider the accidental condition under which this matter has 
combined, it is just possible that some new molecular arrangement 
or combination of elements may have taken place. It is very 
probable that many of these polished balls are much more complex 
in their elementary constitution than I have stated. They are in 
fact a kind of glass, and many of them merely bulbs. Pelouze 
states that glass is probably an indefinite mixture of definite sili- 
cates. Glass, containing small quantities of ferrous oxides and 
sodic sulphates, when exposed to sunlight becomes yellow, and 
possibly some of these balls may have changed in colour since they 
came from the flue. Hydrochloric and nitric acid exert very little 
action on the ferruginous globes: this may be due in some mea- 
sure to the high temperature at which the oxides have been 
formed ; in other cases they are no doubt protected by an external 
coating of some silicate. It would require much time and patience 
to collect a sufficient number of each kind of these minute objects 
for a chemical analysis; but the spectroscope might probably 
assist in revealing their constitution. When time permits I hope 
to resume the subject. 
