CHEMICAL  ANALYSES  BY  SPECTRUM  OBSERVATIONS.  227 
the  composition  of  the  gases  of  the  flame,  the  surprising  result 
came  out  that  the  eye  is  able  to  detect  with  the  greatest  ease 
Quantities  of  sodium  salt  less  than  I      of  a  milliVramme  in 
^  3000000  ° 
weight.  The  reaction  of  potassium  is  not  nearly  so  delicate  ; 
the  spectrum  contains  only  two  characteristic  lines,  one  in  the 
outermost  red,  and  the  other  far  in  the  violet  ray  of  the  solar 
spectrum — points  at  which  the  eye  ceases  to  be  sensitive  to  the 
rays.  The  presence,  however,  of  J_  of  a  milligramme  of  the 
metal  could  be  readily  detected.  Lithium  gives  two  sharply 
defined  lines — the  one  a  very  weak  yellow  line,  and  the  other  a 
bright  red  line,  both  towards  the  extreme  red  end  of  the  solar 
spectrum;  though  the  reaction  is  not  so  sensitive  as  with  sodium, 
it  is  by  far  the  most  delicate  test  for  the  metal,  the  eye  being 
capable  of  distinguishing  with  absolute  certainty  a  quantity  of 
carbonate  of  lithium  less  than  ^      of  a  milli^jramme  in  weight. 
lOUOOOOO  °  o 
The  authors  found  to  their  surprise  that  lithium,  instead  of  be- 
ing a  rare  substance,  was  a  very  widely-distributed  one,  occur- 
ring in  almost  all  bodies.  They  found  it  in  the  water  of  the 
Atlantic;  in  the  ashes  of  marine  plants;  in  pure  spring  water ; 
in  the  ashes  of  tobacTco,  vine  leaves,  and  of  grapes  ;  and  even  in 
the  milk  of  animals  fed  on  crops  growing  in  the  Rhine  plain,  on 
a  non-granite  soil.  Strontium,  barium,  and  calcium  all  give 
characteristic  spectra  ;  that  of  strontium  is  characterized  by  the 
absence  of  green  bands.  It  contains,  however,  eight  remarkable 
ones,  namely,  six  red,  one  orange,  and  one  blue  line.  To  examine 
the  intensity  of  the  reaction,  KirchofF and  Bunsen  threw  up  into  the 
air  of  the  room,  in  the  form  of  fine  dust,  0-077  grm.  of  chloride, 
and  thoroughly  mixed  the  air  by  rapidly  moving  an  umbrella  ; 
the  line  immediately  came  out,  and  realized  the  presence  of  the 
^  part  of  a  milligramme  of  strontium.  The  barium  spectrum 
is  distinguished  by  two  very  distinct  green  lines,  by  which  the 
authors  were  enabled  to  detect  with  certainty  J_  of  a  milli- 
^  1000 
gramme  of  the  metal.  Calcium  gives  a  broad  and  very  charac- 
teristic green  line,  and,  moreover,  a  bright  orange  line  lying 
near  the  red  end  of  the  spectrum.  ^-^^^^  of  a  milligramme  of 
the  chloride  of  the  metal  could  be  easily  detected.  It  is  par- 
ticularly worthy  of  note  that  the  spectra-reactions  of  different 
metals  do  not  interfere  with  one  another ;  that  each  being  cha- 
