A. W. Wright—Spectrum of the Zodiacal Light. 39 
in twist of the same, would, in the width of the grating, equal 
the distance between two individual lines. Take, for instance, 
the 2000 line Nobert, the errors of which are pointed out in 
fig. 8. It would seem that to treble the number of lines 
would, in this case, pretty much destroy the effect of the grat- 
ing ; since the two errors remain independent of the number of 
lines. 
lines of known value; as in Nobert’s microscopic test plate. 
Of course, it is to be expected that the higher powers would be 
dispensed with ; a much lower power being used than would 
be required to separate the lines themselves. 
New Haven, Conn., March, 1874. 
Art. VI.—On the Spectrum of the Zodiacal Light; by ARTHUR 
W. WRieHr. 
THE observations, of which an account is here given, were 
made at various times during the year past, wit _a@ view to 
determine the nature of the zodiacal light, so far as it could be 
effected by a study of its spectrum, and as supplementary to 
the investigations upon its polarization published in this 
ournal in May last. Certain of the statements there made 
were based upon evidence derived from these observations, 
which it is the purpose of this article to set forth more fully. 
As the object studied is one of the faintest among those 
upon which the spectroscope may be employed, some modifica- 
tions were found necessary. both in the instrument and in the 
mode of observation. A Duboscq spectroscope with a single 
visible would greatly weaken, if not obliterate, the faint spec- 
