156 
Intelligence  and  Miscellaneous  Articles. 
The  light  from  the  hydrogen-tube  passed  through  a  lens  which  con- 
centrates it  on  the  slit  of  the  spectroscope.  A  dispersive  power  of 
four  prisms  of  60°  was  used,  the  arrangement  of  the  instrument  being 
such  that  the  ray  of  light  traverses  each  prism  twice.  The  room  is 
kept  perfectly  dark,  and  sunlight  is  reflected  down  from  the  roof  by 
means  of  a  heliostat.  At  first  I  tried  the  usual  mode  of  comparing 
spectra,  viz.  by  having  the  hydrogen  and  solar  spectra  side  by  side. 
I  found  this  answer  very  well  for  the  bright  lines  ;  but  the  faint  ones 
could  not  be  distinguished  by  the  side  of  the  bright  solar  spectrum. 
I  therefore  placed  a  very  fine  platinum  wire  in  the  eyepiece  of  the 
spectroscope,  and  brought  the  lines  under  examination  into  coinci- 
dence with  the  wire,  and  then  passed  the  sunlight  in,  and  found 
which  black  line  coincided  with  the  wire,  or,  where  there  was  no 
coincidence,  the  position  of  the  wire  with  respect  to  the  black  lines. 
I  have  made  from  ten  to  twenty  observations  on  each  line  that  I 
have  at  present  examined.  I  have  every  reason  to  believe  that  the 
limit  of  error  is  within  two  divisions  of  Fraunhofer's  scale  either 
way.  The  Table  below  gives  the  positions  of  the  lines  I  have  already 
compared.  These  I  hope  to  examine  again  next  year,  and  also  to 
finish  the  remainder. 
Reference 
Position  on 
Relative 
Number. 
Kirchhoff's 
brightness. 
Remarks. 
scale. 
10=brightest. 
1* 
694 
10 
c. 
2. 
881 
Limit  of  a  number  of  close  lines  towards  C. 
3. 
930 
5 
Brightest  red  line,  except  C. 
4. 
1014 
5 
Suspiciously  nearthe  chromosphere  line  near  D. 
5. 
1049 
3 
f  The  positions  of  these  were  taken  by  refer- 
6. 
1061 
4 
\      ence  to  the  mercury  lines,  and  are  there- 
7. 
1119 
3 
fore  not  so  reliable  as  the  others. 
8. 
1533 
4 
9. 
1621 
4 
10. 
1876 
4 
11. 
1943 
3 
12. 
1991 
6 
13. 
2065 
3 
14. 
2080 
10 
F. 
15. 
16. 
2235 
2361 
4 
6 
Near  here,  exact  place  uncertain. 
17. 
18. 
19. 
2428-5 
2540 
2605 
2 
Limit  of  a  band  towards  F. 
3 
20. 
2670 
3 
21. 
2767 
Faint  band. 
On  comparing  the  above  Table  with  the  catalogue  of  chromo- 
spheric  lines  by  Professor  Young,  published  in  the  Philosophical 
Magazine  of  November  1871, 1  see  no  sufficient  signs  of  coincidence 
to  lead  me  to  believe  that  any  of  the  chromospheric  lines  in  his  list 
are  due  to  hydrogen,  except  the  C  and  F,  already  well  known  to  be 
so  due.  Since  I  have  not  yet  examined  lines  further  than  2767,  the 
"  near  G  "  (2796)  and  h  lines,  also  known  to  be  due  to  hydrogen,  are 
