ON  THE  PEAK  OF  TENEEIEFE. 
491 
The  brightness  stated  is  that  of  the  brightest  part  of  the  whole  mass,  which  was  usually 
some  20°  to  30°  below  the  apex.  This  point  was  rounded  otf,  and  the  sides  leading  up 
to  it  were  convex  on  the  outside,  in  as  far  as  one  can  speak  certainly  of  such  faint  light ; 
and  tended  much,  by  the  regular  and  complete  symmetrical  form  which  they  indicated, 
to  distinguish  it  from  the  admixture  of  the  Milky  Way  which  crossed  it.  1 could,  how- 
ever, at  times  fancy  that  there  was  a sort  of  wisp  of  the  last  degree  of  faintness  stretch- 
ing fi’om  the  apex  some  50°  further  across  the  sky ; and  on  one  or  two  occasions  could 
almost  persuade  myself  that  it  stretched  all  across  the  sky  to  the  opposite  horizon,  but 
could  never  satisfy  myself  that  it  was  not  fancy. 
Kespecting  alleged  observations  of  the  zodiacal  light  at  midnight,  the  result  of  Gua- 
jara  and  Alta  Vista  is,  that  it  was  decidedly  not  visible  there  either  east  or  west  at  that 
time  (see  August  7 and  September  4) ; ^.  e.  nothing  certain,  nothing  that  seemed  either 
worth  obsening  or  even  possible  to  obsene ; and  assuredly  nothing  approaching  the 
visibility  of  even  the  faintest  part  of  the  area,  included  in  the  ideal  outline  drawn  each 
night  on  the  sky,  when  it  was  seen  E.  or  W.,  so  as  to  include  everything  that  could  be  at  all 
acknowledged  for  zodiacal  light.  Westward  it  was  not  seen  in  the  evening  within  three 
hours  of  midnight ; but  eastward,  and  on  the  high  elevation  of  Alta  Vista,  something 
was  just  visible  at  about  1 a.m.  ; a ditference  explainable  by  the  ah’eady  stated  numerical 
results,  where  the  length  from  the  sun  in  the  eastern  sky  is  given  at  13°  longer  than  in 
the  western,  and  the  brightness  at  twelve  times  greater.  The  midnight  observations 
may  also  be  considered  conformable  to  the  absolute  solar  lengths  found  at  other  times, 
viz.  71°  and  84°. 
If,  then,  at  midnight  nothing  was  seen,  and  at  the  best  period  of  visibility  afterwards, 
■viz.  from  3 to  4 a.ji.,  the  greatest  measm-ed  solar  length  of  any  acknowledged  portion  of 
the  zodiacal  light  was  84°,  it  follows  that  a glow  occasionally  seen  in  the  western  sky  at 
that  morning  horn-,  could  not  be  the  other  end  of  the  zodiacal  light ; for  that  would 
imply  a length  in  the  end,  preVously  found  the  shorter,  of  upwards  of  170°.  The  glow 
in  the  west,  then,  during  the  morning  exhibition  of  the  zodiacal  light  in  the  east,  is  but  a 
reflexion  of  the  latter  on  the  atmosphere,  and  was  closely  paralleled  on  several  occasions 
by  similar  reflexions  of  lunar  dawn. 
On  August  18th  there  was  observed  with  the  lunar  dawn  a cucumstance  that  looked 
at  first  very  much  like  a lunar  zodiacal  light.  At  17**  5“  the  moon’s  twilight  was  visible 
as  a low  flat  elliptical  arch  of  faint  light ; at  17**  12”  it  had  manifestly  grown  pyramidal 
or  pointed  above;  at  17^  15“  the  point  had  extended  itself  mto  a cone,  all  of  the  faint- 
est light,  30°  high  and  some  12°  broad;  at  17‘' 20“  the  moon  rose.  The  cone  looked 
exceedingly  like  a limar  zodiacal  light ; but  on  measuring  its  angle  "with  the  horizon  and 
finding  it  always  90°,  while  the  angle  of  the  ecliptic  at  that  part  of  the  sky  was  only 
38°,  the  appearance  was  manifestly  a mere  local  phenomenon  of  lunar  dawn. 
The  change  from  the  flat  arch  to  the  pointed  cone  is  to  be  observed  in  the  solar  dawn 
as  well ; and  with  the  latter,  as  seen  from  the  mountain,  accompanied  by  overpowering 
light  and  brilliant  colours,  is  precisely  the  feature  which  used  to  make  the  measure- 
