324 Scientific Intelligence. 



nickel, and oopper, a Bel f -induction coil being interposed to elimi- 

 nate the air spectrum and the short metallic lines. * * Of seven 

 plates of spectra photographed in the months of April and May 

 two were very carefully measured with the micrometer. Twenty- 

 two lines were positively identified with those of elements known 

 to be contained in atmospheric dust. The lines are principally 

 the ultimate lines of the respective elements, viz. of calcium, 

 lead, carbon, iron, manganese, nickel, copper, magnesium. The 

 calcium lines were prominent and the presence of calcium carbon- 

 ate in very minute particles of dust suspended in the air was 

 definitely proven. Copper was also prominent and its prevalence 

 in the dust was accounted for by the repeated flashes on the over- 

 head cahle, which is in the center of a wide street (St. Stephen's 

 Green) at a distance of 54 feet from the window of the room in 

 which the experiments were made. The condensation of the 

 vapor of copper after each discharge would yield a dust of extreme 

 tenuity such as could not arise from mechanical action. 



The author adds that " Determinations of the weight of mate- 

 rial necessary to give the lines of these spectra in the manner 

 described, and with such short exposures, have been completed. 

 From the loss of weight of the dry electrodes by the passage of 

 the spark during successive intervals of 10 minutes, the weight 

 of metallic calcium volatalized in one minute, and in one second, 

 was calculated. The number of spark discharges per second was 

 accurately determined, and likewise the number of discharges 

 necessary to render the lines of calcium, as photographed from 

 atmospheric dust, with an exposure 60 seconds. Similarly with 

 copper, the number of discharges necessary to render the ultimate 

 line and the two lines respectively, as measured on plates 

 exposed to atmospheric dust for one second and for five seconds, 

 was ascertained, and the weight of copper volatilized at each 

 spark discharge was accurately determined. It was found that, 

 in order to render the five principal lines in the spectrum of 

 calcium, the weight of the metal passing between the electrodes 

 in 60 seconds was from 00001 to 0-00014 mgrm., or, as calcium car- 

 bonate, from 000025 to 0-00035 mgrm. Similarly, the quantity 

 of copper which passed between the electrodes in one second 

 was from 0-0005 to - 0007 mgrm., and in five seconds from 

 0001 to - 0014 mgrm. These quantities yield the spectra photo- 

 graphed, and the proportion of copper in the dust thus appears 

 to be 10 times as great as that of the calcium. 



" Both the calcium and copper reactions in the spark are more 

 delicate than the sodium test by the yellow flame, or even by the 

 photography of the ultimate lines of sodium in the oxyhydrogen 

 flame and spark. Metallic sodium renders no lines with one spark 

 when photographed, and with five sparks the lines A. 3301-1 and 

 A 3302-5 are both stronger than the yellow lines (mean A 5893 - 2). 

 These lines do not appear on any of the cadmium plates. 



" The reactions of lead, manganese, and magnesium in the spark 

 are much more sensitive than those of sodium, calcium, or copper. 



