258 CARNEGIE INSTITUTION OF WASHINGTON. 



The photographs furnish material for the usual classification of lines 

 according to the temperatures at which they appear and their rate of 

 increase in intensity with rise of temperature. Certain lines are thus 

 picked out which are useful as temperature indicators in the spectra 

 of other sources, lines which are strong at the lowest furnace tem- 

 peratures being of special interest. The fact that a considerable pro- 

 portion of the hnes in these spectra are known to be connected by 

 series relations gives an opportunity to note the behavior of such lines 

 at various temperatures. While the members of the same are afTected 

 similarly, large dilTerences are found between separate series. The 

 members of a triplet series of calcium were tested for a change of rela- 

 tive intensity with wave-length at difTerent temperatures. As the tem- 

 perature rose, the violet members of the series were strengthened to 

 a greater degree than those of longer wave-length. This agrees with 

 furnace observations of the principal series of caesium and indicates a 

 shift of the maximum with change of temperature similar to that shown 

 by the spectrum of an incandescent solid. The close resemblance 

 in structure between the spectra of calcium, strontium, and barium 

 permits a comparison of the homologous lines of the three elements. 

 Such lines are found to be afTected in the same way in the furnace spectra. 



Study of Bandkd Spectra. 



The fluted spectra occurring in sun-spots wliich are ascribed to tita- 

 nium oxide, magnesium hydride, and calcium hydride have been studied 

 to observe the effect of oxygen and hydrogen in producing these bands 

 in the furnace. The results showed clearly that the first-named 

 spectrum is due to titanium oxide, the strength of the bands being con- 

 trolled by the volume of oxygen which was passed directly into the 

 furnace-tube. A sufficient supply of oxygen resulted in a complete 

 suppression of the line spectrum, thus indicating that titanium, when 

 fully changed to the oxide, gives the band spectrum alone. The violet 

 members of the series of titanium oxide bands were found to become 

 relatively stronger as the temperature rose, as in the case of series in hne 

 spectra. The use of hydrogen in the furnace with magnesium and with 

 calcium strengthened the bands in each case, though apparently not in 

 proportion to the supply of the gas, a small quantity of hydrogen, 

 together with low temperature, proving veiy effective in producing the 

 bands. 



Sl'ECTllA OF THK " FuRNACE-FlAME." 



Passing a stream of oxygen through the highly heated furnace-tube 

 affords a means of producing through a wide range of temperature 

 those effects usually obtained in flames by the ignition of various 

 gases in the presence of oxygen. In addition to the results for tita- 

 nium which have been noted, an investigation of the iron spectrum is 

 in progi'ess. The preliminary results indicate that the method will 



