Constituents of Dust and Soot from various Sources. 99 



grey and it was magnetic. It contained iron, copper, and sodium, 

 lead, calcium, potassium, manganese, nickel, silver, thallium a trace, 

 gallium and rubidium a trace, doubtful. 



(III.) Pumice from Krakatoa eruption 1883; from Professor O'Reilly. 

 By decomposing the silicate with ammonium fluoride and sulphuric 

 acid, and precipitating the solution with ammonia, the following bases 

 were separated : iron, copper, silver, sodium, nickel, potassium, rubi- 

 dium, manganese, gallium, and indium a trace.* 



The salt separated by filtration and evaporation of the filtrate 

 contained sodium, potassium, calcium, copper, silver, strontium, nickel 

 a trace, rubidium, and manganese. With the very notable exceptions 

 of strontium, nickel, and cobalt we have found these constituents in 

 ninety-seven irons, ores and associated minerals.! On the other hand, 

 in the examination of six meteoric irons, we have found the same ele- 

 ments invariably associated with nickel and cobalt, the last-named being 

 always in much smaller proportion than the nickel. % Had it been 

 possible to operate on larger quantities, we quite expect that cobalt 

 would have been found in this dust, but the small amount of 8 centi- 

 grams is insufficient for such a purpose, even in the case of most 

 meteoric irons. It is rather a striking fact that in the dust No. 2 

 there is a trace of thallium. This is rather suggestive of its being 

 probably pyrites flue dust, a substance which might occur in hail or 

 rain in a neighbourhood where sulphuric acid is manufactured. It 

 might possibly come from an admixture of soot yielded by a coal con- 

 taining thalliferous pyrites. 



There are three vitriol works within 2 or 3 miles of the College, 

 but after taking all the facts into consideration, we are not able to 

 admit this source as a probable means of contamination, for as will 

 be seen from analyses to be presented, there is one notable constituent 

 we have found in flue dust which is absent from the samples I and 11^ 

 namely, indium. 



In 1897, in order to push this inquiry somewhat further, dust was 

 collected in porcelain dishes placed upon a grass plot in the garden of 

 a residence just on the outskirts of Dublin§ during a period from 

 the 15th November to the 15th December. A considerable fall of a 

 carbonaceous-looking matter occurred on the 16th and 17th of No- 

 vember ; some of the particles were 2 or 3 mm. in diameter, and had a 

 steel grey appearance rather like hard coke or graphite. These 

 particles all sank in the rain-water which collected on the 17th or 

 18th, while a large number of sooty particles floated ; as the dish 

 became over-filled, the sooty matter was automatically washed away 



* 'Trans. Chem. Soc.,' vol. 79, p. 61, 1901. 



t Nickel was found in twenty-three. ' Trans. Chem. Soc.,' vol. 71, p. 533, 1897. 

 X ' Sci. Proc. Dublin Soc.,' New Series, vol. 8. 



§ At the back of my house and remote from any factory chimneys. — W. N. H. 



