* OP ARTS AND SCIENCES. 167 



sliould happen at a time when the azimnth of the pole of the ecliptic 

 is approximately equal to the magnetic declination, we might have an 

 auroral baud along the zodiac which would be taken for zodiacal light. 

 It is mainly for this reason that it has seemed desirable to provide in 

 the Appendix for approximate determinations of the azimuth of the 

 pole of the ecliptic. 



Eylert's observations of December 8 and 9, 1873 (H. 49, 50), afford 

 instances where the presence of auroral light may be suspected. On 

 these dates he saw a luminous arch crossing the heavens, the western 

 part of which had the customary form and position of zodiacal light ; 

 but the eastern part made an angle of about 17° with the ecliptic, from 

 which it deviated to the southeast. The observer was in latitude 

 about +20°, longitude about 30° west of Greenwicli, where the mag- 

 netic declination is about 15° west. The azimuth of the north pole 

 of the ecliptic was about 23° west of north, and its zenith distance 

 about 7G° for the mean of the two observations. It follows that the 

 aberrant portion of the zodiacal light was more nearly perpendicular 

 to the magnetic meridian than to the circles of latitude, but tlie devi- 

 ation was in excess of what would be required on the hypothesis that 

 an auroral band was seen. It may not be always the case that these 

 auroral bauds are perpendicular to the magnetic meridian, and the 

 evidence of auroral light in this instance is not at all conclusive. 



A few days later, on the evening of December 12, 1873, Serpieri, 

 at Urbino in Italy, observed a general illumination of the zodiac 

 (II. 51). The occurrence of coruscations in the light on this occasion 

 perhaps adds a little weight to the suspicion of auroral action, but these 

 coruscations, or pulsations, as Jones called them, have often been seen 

 in the zodiacal light proper. The nature of this phenomenon, also, 

 evidently requires photometric investigation. On the evening of 

 January 31, 1883, I found, during half an hour's observation, that 

 there were apparent variations in the relative brightness of definite 

 portions of the zodiacal light with respect to definite portions of the 

 Milky Way. But the variation may not have been real, or it may 

 possibly have been due to variable haze, too indefinite in outline to 

 be recognized. 



It is not my intention, on this occasion, to discuss the probability of 

 any explanation of the zodiacal light. I have merely to remark, with 

 regard to the ordinary meteoric theory, that it gains greatly in sim- 

 plicity if we dispense with all the imaginary meteoric bodies or rings 

 with which it has usually been connected, and retain merelv the con- 

 ception of meteoric dust difTused throughout the Solar System. It may 



