442 MR. THOMAS HEELIs's OBSERVATIONS 



at tlie apex much beyond it_, were first observed. I find 

 the observation recorded in my note-book in the following 

 termSj written down at the time^ and which may be taken 

 to be a fair description of the peculiarities noted in the 

 observations of the faint envelope : — ^' There seems a fainter 

 kind of luminous envelope surrounding the true lights as 

 if it were cigar-shaped^ in layers. This outer envelope is less 

 bright than the inner [cone] , especially near the apex, and 

 near the horizon it tones ofi" into the other, thus accounting 

 for the large breadth assigned to the light near the horizon. 

 At the base the two envelopes are undistinguishably mixed. 

 I have often noticed this before, but have not included the 

 envelope in my measures in cases in which I could dis- 

 tinguish it from the true light. The luminous envelope 

 this evening was, at Prsesepe, about two- thirds of the bright- 

 ness of the true light.^^ The light of this envelope was by 

 no means equable. It seldom appeared at all until long 

 after the departure of the twilight permitted observations 

 of the main body of the light, and it generally made its 

 appearance as a faint streak, most frequently on the southern 

 side of the light, and extending from the limb near the 

 apex, apparently overlapping the limb, and extending far 

 beyond the apex. As the night wore on, and the main 

 body of the light sank beneath the horizon, the other limb 

 of the envelope appeared, completing the cone. The main 

 body of the light was hardly ever, near the apex, shaded 

 gradually into the envelope; but the space between the 

 brightest part of the latter and the main body of light was 

 comparatively dark, and gave me the impression of looking 

 into space through a very thin crape veil. I was never able 

 to separate the envelope distinctly from the main body of 

 the light in the few observations which I made in the 

 morning. 



The following are the observed lengths of the envelope. 

 The Table is arranged in the same way as that in which 



